A comparison of document clustering techniques

A comparison of document clustering techniques (2000)

Cached

Download Links

by Michael Steinbach , George Karypis , Vipin Kumar

Venue:	In KDD Workshop on Text Mining
Citations:	454 - 27 self

BibTeX

@INPROCEEDINGS{Steinbach00acomparison,
    author = {Michael Steinbach and George Karypis and Vipin Kumar},
    title = {A comparison of document clustering techniques},
    booktitle = {In KDD Workshop on Text Mining},
    year = {2000}
}

Years of Citing Articles

Bookmark

OpenURL

Abstract

This paper presents the results of an experimental study of some common document clustering techniques: agglomerative hierarchical clustering and K-means. (We used both a “standard” K-means algorithm and a “bisecting ” K-means algorithm.) Our results indicate that the bisecting K-means technique is better than the standard K-means approach and (somewhat surprisingly) as good or better than the hierarchical approaches that we tested.

Citations

6858	A mathematical theory of communication - Shannon - 1948 (Show Context) Citation Context ...l the clustering is working by comparing the groups produced by the clustering techniques to known classes. This type of measure is called an external quality measure. One external measure is entropy =-=[Sha48]-=-, which provides a measure of “goodness” for un-nested clusters or for the clusters at one level of a hierarchical clustering. Another external measure is the F-measure, which, as we use it here, is m...
2290	Dubes. Algorithms for Clustering Data - Jain, C - 1988 (Show Context) Citation Context ...d or better than those produced by agglomerative hierarchical clustering techniques. We have also been able to find what we think is a reasonable explanation for this behavior. We refer the reader to =-=[2]-=- for a review of cluster analysis and to [4] for a review of information retrieval. For a more complete version of this paper, please see [6]. The data sets that we used are ones that are described mo...
1480	Finding Groups in Data: An Introduction to Cluster Analysis - Kaufman, Rousseeuw
674	Scatter/Gather: A cluster-based approach to browsing large document collections - Cutting, Karger, et al. - 1993 (Show Context) Citation Context ...ught to produce inferior clusters. Sometimes K-means and agglomerative hierarchical approaches are combined so as to “get the best of both worlds.” For example, in the document domain, Scatter/Gather =-=[1]-=-, a document browsing system based on clustering, uses a hybrid approach involving both K-means and agglomerative hierarchical clustering. K-means is used because of its run-time efficiency and agglom...
444	Hierarchically classifying documents using very few words - Koller, Sahami - 1997 (Show Context) Citation Context ...2] or in organizing the results returned by a search engine in response to a user’s query [ZEMK97]. Document clustering has also been used to automatically generate hierarchical clusters of documents =-=[KS97]-=-. (The automatic generation of a taxonomy of Web documents like that provided by Yahoo! 1(www.yahoo.com) is often cited as a goal.) A somewhat different approach [AGY99] finds the natural clusters in...
353	A robust clustering algorithm for categorical attributes - GUHA, RASTOGI, et al. - 1999 (Show Context) Citation Context ...e fixed once they happen. In cases where nearest neighbors are unreliable, a different approach is needed that relies on more global properties. (This issue was discussed in a non-document context in =-=[3]-=-.) Since computing the cosine similarity of a document to a cluster centroid is the same as computing the average similarity of the document to all the cluster’s documents [6], K-means is implicitly m...
243	Refining Initial Points for k-Means Clustering - Bradley, Fayyad - 1998
158	Fast and effective text mining using linear-time document clustering - Larsen, Aone - 1999 (Show Context) Citation Context ...e clusters at one level of a hierarchical clustering, and the F-measure, which measures the effectiveness of a hierarchical clustering. (The F measure was recently extended to document hierarchies in =-=[5]-=-.) Our results will show that the bisecting K-means algorithm has the best performance for these two measures of cluster quality. 3. Bisecting K-means The bisecting K-means algorithm starts with a sin...
128	Information Retrieval Systems. Theory and Implementation, 1st edn - Kowalski - 1997 (Show Context) Citation Context ...tive hierarchical clustering techniques. We have also been able to find what we think is a reasonable explanation for this behavior. We refer the reader to [2] for a review of cluster analysis and to =-=[4]-=- for a review of information retrieval. For a more complete version of this paper, please see [6]. The data sets that we used are ones that are described more fully in [6]. They are summarized in the ...
104	Optimization of inverted vector searches - Buckley, Lewit - 1985
103	Fast and intuitive clustering of web documents - Zamir, Etzioni, et al. - 1997 (Show Context) Citation Context ...ument [BL85]. More recently, clustering has been proposed for use in browsing a collection of documents [CKPT92] or in organizing the results returned by a search engine in response to a user’s query =-=[ZEMK97]-=-. Document clustering has also been used to automatically generate hierarchical clusters of documents [KS97]. (The automatic generation of a taxonomy of Web documents like that provided by Yahoo! 1(w...
74	WebACE: A web agent for document categorization and exploration - Han, Boley, et al. - 1998
47	On the Merits of Building Categorization Systems by Supervised Clustering - Aggarwal, Gates, et al. - 1999 (Show Context) Citation Context ...archical clusters of documents [KS97]. (The automatic generation of a taxonomy of Web documents like that provided by Yahoo! 1(www.yahoo.com) is often cited as a goal.) A somewhat different approach =-=[AGY99]-=- finds the natural clusters in an already existing document taxonomy (Yahoo!), and then uses these clusters to produce an effective document classifier for new documents. Agglomerative hierarchical cl...
38	Comparison of hierarchic agglomerative clustering methods for document retrieval - El-Hamdouchi, Willett - 1989 (Show Context) Citation Context ...ecide, at each step, which cluster to split and how to perform the split. Agglomerative techniques are more common, and these are the techniques that we will compare to K-means and its variants. (See =-=[EW89]-=- for a comparison of agglomerative hierarchical clustering methods used for document retrieval.) We summarize the traditional agglomerative hierarchical clustering procedure as follows: Simple Agglome...
38	Chinatsu Aone. Fast and effective text mining using linear-time document clustering - Larsen - 1999 (Show Context) Citation Context ...agglomerative hierarchical clustering. K-means is used because of its efficiency and agglomerative hierarchical clustering is used because of its quality. Recent work to generate document hierarchies =-=[LA99]-=- uses some of the clustering techniques from [CKPT92] and presents a result that indicates that agglomerative hierarchical clustering is better than Kmeans, although this result is just for a single d...
31	The star clustering algorithm for static and dynamic information organization - Aslam, Pelekov, et al. - 2004
20	A Comparison of Document Clustering - Steinbach, Karypis, et al. (Show Context) Citation Context ...able explanation for this behavior. We refer the reader to [2] for a review of cluster analysis and to [4] for a review of information retrieval. For a more complete version of this paper, please see =-=[6]-=-. The data sets that we used are ones that are described more fully in [6]. They are summarized in the following table. 1 Table 1: Summary description of document sets. Data Set Source Documents Class...
8	Reuters-21578 text categorization text collection 1.0 - Lewis - 2004 (Show Context) Citation Context ...es of articles, such as “Entertainment”, “Financial”, “Foreign”, “Metro”, “National”, and “Sports.” Data sets re0 and re1 are from Reuters-21578 text categorization test 9collection Distribution 1.0 =-=[reut]-=-. Data set wap is from the WebACE project (WAP) [han98]. Each document corresponds to a web page listed in the subject hierarchy of Yahoo!. Data Set Source Number of Documents Number of Classes Minimu...
5	Dubes and Anil K. Jain, Algorithms for Clustering Data - Richard - 1988 (Show Context) Citation Context ...ering techniques that are commonly used for document clustering. Agglomerative hierarchical clustering is often portrayed as “better” than K-means, although slower. A widely known study, discussed in =-=[DJ88]-=-, indicated that agglomerative hierarchical clustering is superior to K-means, although we stress that these results were with non-document data. In the document domain, Scatter/Gather [CKPT92], a doc...
1	Optimizations of inverted vector searches, SIGIR ’85, Pages 97110 - Buckley, Lewit - 1985 (Show Context) Citation Context ...ally, document clustering was investigated for improving the precision or recall in information retrieval systems [Rij79, Kow97] and as an efficient way of finding the nearest neighbors of a document =-=[BL85]-=-. More recently, clustering has been proposed for use in browsing a collection of documents [CKPT92] or in organizing the results returned by a search engine in response to a user’s query [ZEMK97]. Do...